LPDi GAN: A License Plate De-Identification Method to Preserve Strong Data Utility
- Xiying Li 1,2,3, Heng Liu 1,2,3, Qunxiong Lin 2, Quanzhong Sun 2, Qianyin Jiang 4, Shuyan Su 1,2,3
- Xiying Li 1,2,3, Heng Liu 1,2,3, Qunxiong Lin 2
- 1School of Intelligent Systems Engineering, Sun Yat-sen University, Shenzhen 518107, China.
- 2Key Laboratory of Video and Image Intelligent Analysis and Application Technology, Ministry of Public Security, Guangzhou 510006, China.
- 3Guangdong Provincial Key Laboratory of Intelligent Transportation System, Shenzhen 518107, China.
- 4School of Computing, Guangzhou Maritime University, Guangzhou 510725, China.
- 0School of Intelligent Systems Engineering, Sun Yat-sen University, Shenzhen 518107, China.
Related Experiment Videos
Contact us if these videos are not relevant.
Contact us if these videos are not relevant.
View abstract on PubMed
Summary
This summary is machine-generated.This study introduces LPDi GAN, a novel method for de-identifying license plates (LPs) in images. The generative adversarial network creates synthetic LPs, preserving data utility and privacy in transportation datasets.
Area Of Science
- Computer Vision
- Artificial Intelligence
- Data Privacy
Background
- License plate (LP) information is sensitive personal data.
- Existing de-identification methods like blurring reduce data utility.
- Public transportation datasets often contain unprocessed LP images.
Purpose Of The Study
- To propose a novel method for license plate de-identification.
- To generate synthetic license plates that maintain data utility.
- To enhance privacy in transportation datasets using generative adversarial networks.
Main Methods
- Developed a generative adversarial network (LPDi GAN) for LP de-identification.
- Extracted background features to generate similar LPs.
- Incorporated LP templates and styles for controllable character generation.
Main Results
- LPDi GAN effectively de-identifies images while preserving data utility.
- The method adapts to environmental conditions and LP tilt angles.
- Achieved a Learned Perceptual Image Patch Similarity (LPIPS) of 0.25, ensuring character recognition.
Conclusions
- LPDi GAN offers superior de-identification compared to traditional methods.
- The approach balances privacy protection with data utility.
- Enables the creation of high-quality synthetic LPs for research and development.
Related Experiment Videos
Contact us if these videos are not relevant.
Contact us if these videos are not relevant.
Related Concept Videos
01:20
When proton-coupled carbon-13 spectra are simplified by a broadband proton decoupling technique, structural information about the coupled protons is lost. Distortionless enhancement by polarization transfer (DEPT) is a technique that provides information on the number of hydrogens attached to each carbon in a molecule. While the DEPT experiment utilizes complex pulse sequences, the pulse delay and flip angle are specifically manipulated. The resulting signals have different phases depending on...
03:31
DNA probes are fragments of DNA labeled with a reporter tag to enable their detection or purification. The resulting labeled DNA probes can then hybridize to target nucleic acid sequences through complementary base-pairing, and may be used to recover or identify these regions.
Radioisotopes, fluorophores, or small molecule binding partners like biotin or digoxigenin, are the most widely used reporter tags for labeling DNA probes. These labels can be attached to the probe DNA molecule via...

